Studies completed at this institution strongly support the hypothesis that monoamine oxidase inhibitors, (MAOI) are superior to tricyclic antidepressants (TCA) and placebo to AD I (atypical depressives). These patients who meet RDC criteria for Major or Intermittent Depressive Disorder are also characterized by reactive mood, weight gain, oversleeping, rejection sensitivity, and lethargy. Post hoc it was observed that AD with a history of panic attacks (without agoraphobia) were least likely to respond to placebo and most likely to benefit from phenelzine, and this accounted for most of the drug vs. placebo variance. Another important unexpected observation was that imipramine, although less effective than phenelzine, is superior to placebo in AD. This suggests that current criteria for AD I include at least four distinct patients subtypes: MAOI responders (the largest group), TCA responders, placebo responders, and nonresponders. A pilot sleep study suggests AD I have the REM abnormality but not the sleep fragmentation observed in most depressives. Pilot data also suggest that patients with only some of the features of the AD syndrome (AD II) also have a superior improvement rate with phenelzine compared to TCA and placebo. We would like to repeat a double-blind trial of phenelzine, imipramine and placebo in AD I and AD II patients which will last 6 weeks for nonresponders and 12 weeks for responders. All patients will also have sleep polysomnographic studies and lactate infusions. Any single study does not prove a scientific fact or establish treatment efficacy. Repetition and extension of the original work will establish the selective utility of MAOI in this patient group. The health related and heuristic implications of this are obvious, and included more effective treatment for outpatient depressives and data supporting the validity of AD as a distinct depressive subtype. Replicating the import of a history of panic attacks will establish a simple clinical marker for identifying patients with antidepressant-sensitive diathesis in a heterogeneous depressed group previously considered "neurotic" by many clinicians. Sleep polysomnographic findings and data from lactate infusions will be used to biologically characterize different types of AD patients (e.g. phenelzine responders, etc.). The work with AD II patients will help establish the boundaries of this syndrome.